Week 8 - Rasterisation and Anti-Aliasing

Question 1

Name 2 examples of vector displays.

Answer 1

Pen plotters
Tektronix cathode ray tubes

Question 2

Name 5 examples of raster displays.

Answer 2

Monitors
Televisions
Laser, ink-jet, dot matrix printers
Liquid crystal displays
Arrays of LEDs

Question 3

What is rasterisation?

Answer 3

The conversion of ideal, mathematical graphics primitives onto raster displays

Question 4

What is scan conversion?

Answer 4

Rasterisation line by line of the raster display

Question 5

How does scanline interpolation work?

Answer 5

Use Z-buffer algorithm to determine screen locations of vertices

Intensities or normals interpolated from vertices

Interpolate along scanline from edges

Question 6

What is the parity rule for filling polygons?

Answer 6

A point is inside a polygon if a line to a distant point intersects boundary an odd number of times

Question 7

What is the non-zero winding number rule for filling polygons?

Answer 7

A point is inside a polygon if the polygon winds clockwise around the point a non-zero number of times

Question 8

What is scan-line coherence?

Answer 8

An interior pixel is adjacent to another interior pixel unless a boundary is encountered

Question 9

How does scan-line coherence work?

Answer 9

For each scan line:
- Find intersections of scan line with all polygon edges
- Sort intersections into increasing order of x coordinate
- Fill alternate segments of scan line (spans) according to parity rule

Question 10

What are the two ways to choose which pixels to illuminate for a line?

Answer 10

Every pixel the line intersects

The vertically closest pixel

Question 11

What algorithm is used to illuminate the pixel vertically closest to a line?

Answer 11

Naïve midpoint algorithm

Question 12

What does DDA stand for?

Answer 12

Digital Differential Analyser

Question 13

What is the Bresenham circle algorithm?

Answer 13

Choose pixel with centre closest to circle

Work on a single octant and draw others by symmetry

Can be written in integer arithmetic

Question 14

What causes aliasing?

Answer 14

Undersampling (not taking enough discrete samples to capture a fast rate of change)

Question 15

What does ADC stand for?

Answer 15

Analog-to-Digital Converter

Question 16

What does DAC stand for?

Answer 16

Digital-to-Analog Converter

Question 17

Why will there always be aliasing, no matter how densely we sample?

Answer 17

Repetitive patterns can have arbitrarily high frequencies due to perspective effects

Object edges create sharp changes -> high frequencies

Question 18

How does anti-aliasing work in computer graphics?

Answer 18

The scene is rendered

No access to a continuous image

Low-pass filtering of continuous image has to be simulated/approximated

Question 19

What are the three filtering approaches?

Answer 19

No filtering

Pre-filtering (simulation)

Post-filtering (approximation)

Question 20

What does SSAA stand for?

Answer 20

Super-Sampling Anti-Aliasing

Question 21

What is SSAA?

Answer 21

Take many point samples per pixel, and assign grey level as the (weighted) average value

Question 22

What is a pro and a con of SSAA?

Answer 22

Pro: smoother, nicer looking results

Con: ridiculously expensive and can lead to low performance and framerate

Question 23

What is multisample anti-aliasing?

Answer 23

Faster processing by processing fragment shaders only once, and only supersampling depth and stencil buffers

Question 24

What does FXAA stand for?

Answer 24

Fast Approximate Anti-Aliasing

Question 25

What is FXAA?

Answer 25

Pixel-level fast algorithm (implemented as fragment shader) identifies depth edges in the image, and smooths edges at pixel level

Question 26

What are two issues to be aware of with anti-aliasing and texture maps, and how are they fixed?

Answer 26

Compression (several texels map to a single pixel) - texture image should be sampled many times to calculate pixels, but this is very expensive

Magnification (single pixel lies deep within texel) - higher resolution texture map required

Question 27

What are the three types of filtering in MIP mapping?

Answer 27

Bilinear filtering, trilinear filtering, anisotropic filtering

Question 28

What is bilinear filtering?

Answer 28

Approximate pixel value from pixels around it

Question 29

What is trilinear filtering?

Answer 29

Also approximate between levels of MIP mapping

Question 30

What is anisotropic filtering?

Answer 30

Sample texture with different horizontal and vertical scale ratios

Question 30

What is anisotropic filtering?

Answer 30

Sample texture with different horizontal and vertical scale ratios